Optimal Investigation as a Regenerative Stopping Problem.
Abstract
This report discuss the problem of dynamic optimal investigation of a two-state (in control, out of control) system. The true state can only be inferred from reported costs and the time since the last correction. It is demonstrated that when the parameters satisfy certain conditions, such problems can be efficiently solved as regenerative stopping problems. Some general results for regenerative stopping problems are also obtained. In the last section the problem is generalized to n two-state systems. By combining simulation with the Regenerative Stopping Algorithm, a problem with 20 state variables is solved with a small error term. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1979
- Accession Number
- ADA078510
Entities
People
- A. Gregory Buckman
- Bruce L. Miller
Organizations
- University of California, Los Angeles